---
license: llama2
library_name: peft
tags:
- trl
- sft
- generated_from_trainer
datasets:
- dialogstudio
base_model: TencentARC/LLaMA-Pro-8B
model-index:
- name: llama-pro-8b-tweet-summarization-gradnorm-0.3
  results: []
---

<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->

# llama-pro-8b-tweet-summarization-gradnorm-0.3

This model is a fine-tuned version of [TencentARC/LLaMA-Pro-8B](https://huggingface.co/TencentARC/LLaMA-Pro-8B) on the dialogstudio dataset.
It achieves the following results on the evaluation set:
- Loss: 2.9796
- Rouge Scores: {'rouge1': 93.71888929189157, 'rouge2': 77.8377567936117, 'rougeL': 64.47906852741538, 'rougeLsum': 93.71298018429633}
- Bleu Scores: [0.9470990193868204, 0.9341779145832757, 0.9064440397746264, 0.8744914403659334]
- Gen Len: 463.0182

## Model description

More information needed

## Intended uses & limitations

More information needed

## Training and evaluation data

More information needed

## Training procedure

### Training hyperparameters

The following hyperparameters were used during training:
- learning_rate: 0.0001
- train_batch_size: 4
- eval_batch_size: 8
- seed: 42
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- num_epochs: 7
- mixed_precision_training: Native AMP

### Training results

| Training Loss | Epoch | Step | Validation Loss | Rouge Scores                                                                                                             | Bleu Scores                                                                      | Gen Len  |
|:-------------:|:-----:|:----:|:---------------:|:------------------------------------------------------------------------------------------------------------------------:|:--------------------------------------------------------------------------------:|:--------:|
| 1.9065        | 1.0   | 220  | 1.8530          | {'rouge1': 92.83694737064799, 'rouge2': 78.72458121869542, 'rougeL': 67.88788283384865, 'rougeLsum': 92.83768512059282}  | [0.8739198483584956, 0.8530170264142946, 0.8271978418182495, 0.7998377773703629] | 463.0182 |
| 1.6363        | 2.0   | 440  | 1.8633          | {'rouge1': 93.54135671371444, 'rouge2': 78.96116387599493, 'rougeL': 67.77857901494997, 'rougeLsum': 93.54432289584433}  | [0.8758125801988195, 0.8577741180618648, 0.8322886881519586, 0.80457236049974]   | 463.0182 |
| 1.2817        | 3.0   | 660  | 2.0098          | {'rouge1': 87.30764070509844, 'rouge2': 73.12328274037898, 'rougeL': 62.00625532521349, 'rougeLsum': 87.29149649901954}  | [0.8757949025917542, 0.8593181834244542, 0.8334473061685955, 0.8048319452251607] | 463.0182 |
| 0.9049        | 4.0   | 880  | 2.2481          | {'rouge1': 87.35996946418575, 'rouge2': 72.87802745947901, 'rougeL': 61.35206821444361, 'rougeLsum': 87.32662841081371}  | [0.8755472589597261, 0.859572654041077, 0.8333237300074641, 0.804082483213136]   | 463.0182 |
| 0.5916        | 5.0   | 1100 | 2.5061          | {'rouge1': 78.38431994557745, 'rouge2': 64.89809559762811, 'rougeL': 53.805209482421525, 'rougeLsum': 78.30608179426231} | [0.747179346815877, 0.7352208249958618, 0.7126103103040894, 0.6869428956670465]  | 463.0182 |
| 0.3898        | 6.0   | 1320 | 2.8150          | {'rouge1': 93.77539618029996, 'rouge2': 78.03050568501187, 'rougeL': 64.82344374456906, 'rougeLsum': 93.76894400818286}  | [0.9469183628254614, 0.9342162110956728, 0.9067374010427977, 0.8750430150656403] | 463.0182 |
| 0.2961        | 7.0   | 1540 | 2.9796          | {'rouge1': 93.71888929189157, 'rouge2': 77.8377567936117, 'rougeL': 64.47906852741538, 'rougeLsum': 93.71298018429633}   | [0.9470990193868204, 0.9341779145832757, 0.9064440397746264, 0.8744914403659334] | 463.0182 |


### Framework versions

- PEFT 0.8.2.dev0
- Transformers 4.38.0.dev0
- Pytorch 2.1.0+cu121
- Datasets 2.16.2.dev0
- Tokenizers 0.15.1